Subscribe to List View Past IssuesRSS translate   facebook facebook Like 0 Comment 0Share twitter
Email not displaying correctly? View it in your browser.


Investment technology for the 21st century.
Second beta of 1.04 released

A new beta version of Portfolio Probe 1.04 is now available.

Getting it for R users

If things are nice for you, then you can do:

install.packages("PortfolioProbeBeta", repos="http://www.portfolioprobe.com/R")

and you will have it installed.

If things are not so nice for you, then you can use a browser to get the zip file from http://www.portfolioprobe.com/R/bin/windows/contrib/

This version is built for R version 2.10 and newer.  If you are using an older version of R, then let us know and we'll do the build.

I had one case where the install.packages command only half-way succeeded (apparently because it was confused about the package already being there) and the code was not available.  But issuing the install.packages command again solved the problem.

Getting it for S+ users

Go to http://www.portfolioprobe.com/Splus/V1.04beta2/ where you will find a dot-q file to source and the DLLs for 32-bit and 64-bit.  If you are blocked from getting a DLL, then download a zip file from the R repository and get the appropriate DLL from the libs directory within the zip file.

Works in R 2.14.x

So why a new beta?  A key reason is that it works with R version 2.14.x.  The problem before was that 2.14 demands a namespace.

Generalized costs

The second key reason was that a user requested a generalization and was anxious to start playing with it.

The cost.par argument can now be a matrix that is the same size as the matrices given for long.buy.cost and its mates.  This allows each asset to have its own market impact model in an easily implementable way.  For example cost.par might be a two-column matrix where each value in the first column is 1 (linear costs) and the values in the second column have a range of maybe 1.5 to 1.8 (market impact costs).

Non-linear costs in long-short

The change above  highlighted that the way non-linear costs were implemented for long-short portfolios in the case that a trade goes from long to short or vice versa was rather silly.  The silliness has been removed, and hence there is a non-backward compatible change in that case.

Other changes

A few minor bugs have been eliminated and a few additions have been made.

Next steps

Before the official release of 1.04, still to be done is:

  • a few code changes
  • more testing
  • documentation of changes
You have been sent this email because you have opted-in on our online request form.

Unsubscribe pburns@pburns.seanet.com from this list.

Our mailing address is:
Burns Statistics
4-b Jodrell Road
London, England E3 2LA

Add us to your address book

Copyright (C) 2012 Burns Statistics All rights reserved.

Forward this email to a friend
Update your profile
Email Marketing Powered by MailChimp